Varicella, Measles, and Mumps

Classification and Morphology

VZV (herpesvirus varicellae) is a member of the Herpesviridae family. In addition to a burgeoning number of animal herpesviruses, this group includes seven additional, closely related viruses that infect humans: herpes simplex viruses (HSV) types 1 and 2 (herpesvirus hominis), cytomegalovirus (CMV), Epstein-Barr virus, and human herpesviruses 6, 7, and 8. Only one antigenic type of VZV has been identified, but molecular studies have revealed some minor differences in VZV that have proved useful for epidemiologic studies.

Common properties of the family include a DNA genome and enveloped virions exhibiting icosahedral symmetry with a diameter of 180 to 200 nm. Nucleocapsids, which are assembled in the nucleus, have a diameter of about 100 nm, and consist of a DNA core surrounded by 162 identical subunits, or capsomeres. Nucleocapsids acquire a temporary envelope at the nuclear membrane; they are transported further via the endoplasmic reticulum to the Golgi, where they receive a final envelope. In cell cultures, virions are packaged in vesicles identified as endosomes, which are acidic. Virus particles are released from these structures at the cell surface by exocytosis. Extracellular virions are extremely pleomorphic compared with virions of HSV. This pleomorphism, presumably reflecting injury to the envelope possibly caused by exposure to acid or enzymes in endosomes, is believed to account for the lability and lack of cell-free virus that characterizes VZV in tissue culture and in its spread through the body during varicella infection, and it distinguishes VZV from HSV. In vivo, enveloped and well-formed VZV is released from cells of the superficial epidermis (strateum corneum), yielding highly infectious virions capable of airborne spread and with a great degree of communicability.

Molecular studies have elucidated details concerning how latent VZV infection is established, maintained, and reactivated. Latent infection comes from virions present in skin during varicella and/or from T-cell associated viremia in varicella. It is unlikely, however, that complete viral replication occurs in the neuron during establishment of latency because the neuron must survive, and replication would be expected to cause cell death. The replication process of VZV is begun during latent infection, but a block in the cascade of viral gene expression probably occurs. Numerous laboratories have found that least six viral genes and their protein gene products are reported to be expressed in latently infected neurons. These proteins are confined to the cell cytoplasm. It seems that when these proteins are transduced into the nucleus by factors still to be determined, reactivation occurs, with formation of all 72 VZV gene products and synthesis of infectious, enveloped virions in tissues such as nerve and skin. Recently, however, a much more restricted gene expression in VZV latency has been proposed. It is currently not yet clear how this controversy regarding VZV latency will eventually be resolved.

Propagation

VZV grows readily in diploid human fibroblasts, such as WI 38 cells, the most commonly used cell type for virus isolation. VZV also can be propagated in certain epithelial cells, such as human embryonic kidney, primary human amnion cells, primary human thyroid cells, and Vero (African green monkey kidney) cells. Similar to CMV, the cytopathic effect of VZV is focal in cell culture because of its cell-associated character, and cytopathic effects develop more slowly (3-7 days) than with HSV. Animal models for varicella (guinea pigs) and for zoster (rats) have been described. An in vitro model of latency and reactivation in guinea pig enteric neurons has also been developed and provides a setting in which to study factors that influence latency and reactivation.

Serologic Tests and Antigenic Properties of Varicella-Zoster Virus

Several serologic tests are available to measure antibodies to VZV, including indirect immunofluorescence, often called fluorescent antibody to membrane antigen (FAMA) ; latex agglutination ; enzyme-linked immunosorbent assay (ELISA) ; radioimmunoassay ; immune adherence hemagglutination ; neutralization ; and complement-enhanced neutralization. All of these methods are more sensitive than the complement fixation assay. Data gathered from these assays show that antibody to VZV develops within a few days after the onset of varicella, persists for many years, and is present before the onset of zoster.

Serologic cross-reactions between HSV and VZV have been described. HSV and VZV share common antigens, and similar polypeptides and glycoproteins have been identified for both viruses, but cross-protection has not been observed. Rare simultaneous infections with one or more human herpesviruses have been reported. Elevations in heterologous antibody titers in apparent HSV or VZV infections may result from cross-reactions of the viruses but also may indicate simultaneous infection by both viruses.

VZV produces at least eight major glycoprotein antigens—B, C, E, H, I, K, L, and M—all of which are on the envelope of the virus and on the surface of infected cells. The glycoproteins and internal antigens, such as the capsid and tegument, stimulate production of neutralizing and other types of antibodies and cellular immunity. The most prominent glycoprotein of VZV is glycoprotein E. Antibodies elaborated in varicella and zoster are of the immunoglobulin G (IgG), IgA, and IgM classes.

Immunity to Varicella-Zoster Virus

Immunity to VZV is a complex interaction between humoral and cell-mediated immune responses, with the possibility of partial and complete immunity to the virus. Humoral and cellular immunity are important in the control of primary varicella infection. Cell-mediated immunity is important for viral clearance, providing long-term protection against varicella and preventing symptomatic VZV reactivation. Immunity to varicella usually is long lasting. However, it might wane occasionally; rarely, second episodes of varicella have been reported among immunocompetent persons.

Clinical reinfection with VZV has been observed in some persons despite a positive antibody titer at exposure. Most clinical reinfections are mild, however, which suggests that partial immunity to the virus may be present. In a study by Bogger-Goren and colleagues, moreover, seronegative children who had positive cellular immune responses were likely to be protected against varicella after household exposure. In contrast, children with negative cellular immune responses became infected. Secretory IgA against VZV has also been demonstrated after varicella. Although it has not been shown, it is further hypothesized that cellular immunity at the mucosal level may play a role in protection against clinical varicella.

Immunologic evidence consistent with asymptomatic reinfection with VZV, manifested by an increase in VZV-specific IgG or IgA or the production of IgM and an increase in the cell-mediated immune response to VZV (external boosting), has been documented in adults with a household exposure to varicella. Subclinical reactivation with possible subsequent boosting of immunity (internal boosting) is also possible.

In addition to predisposing to reinfection with the virus, incomplete immunity to VZV is associated with development of zoster. In addition to clinical zoster, silent reactivation of latent VZV in persons who have had previous varicella probably occurs; this may be detected immunologically by an increase in antibody titer or the transient appearance of specific IgM, although it is difficult to rule out the possibility of an exogenous exposure. Sometimes, clinical manifestations of zoster, such as pain, may occur in the absence of a rash—so-called zoster sine herpete. Silent reactivation of VZV in bone marrow transplant patients has been shown by polymerase chain reaction (PCR) assay. Zoster results in patients who have latent VZV infection when specific cell-mediated immunity is depressed. Defective antibody responses to VZV glycoproteins have not been associated with development of zoster in immunocompromised persons. Similarly, the increased incidence of zoster in elderly adults has been associated with loss of cell-mediated immunity to VZV, whereas antibody to VZV does not wane with age but tends to increase.

It is possible to provide humoral immunity to persons at high risk for developing severe varicella by passive immunization. Although used successfully to prevent severe varicella, passive immunization has not prevented zoster in persons at high risk for it. It is uncertain whether passive immunization of a woman with varicella can prevent infection of her fetus or development of congenital varicella syndrome, although some evidence suggested a benefit to the fetus too. It is possible to increase cell-mediated immunity to VZV by immunization, and this approach was demonstrated in several studies. Results of a large, double-blind, controlled study in healthy vaccinees older than 60 years indicated that approximately half of more than 15,000 vaccinated individuals were protected from developing zoster.

Communicability

Patients with varicella are considered to be contagious for 1 to 2 days before rash becomes obvious and for as long as new lesions continue to appear and while they are moist (usually 5-7 days after rash onset). Historically, transfer of VZV was believed to occur via respiratory droplets, and limited epidemiologic evidence suggests that transmission can occur before the onset of rash although the question of whether the early onset with few localized lesions went unnoticed is appropriate. It is rare, however, to isolate VZV from the pharynx of infected patients, and isolation has occurred only after the onset of rash. A study using PCR methods showed that VZV DNA is present in the nasopharynx of a high percentage of children during the early stages of clinical varicella, but the presence of VZV DNA does not necessarily indicate the presence of infectious virus. In contrast, VZV can readily be cultured from the vesicular lesions in varicella and zoster. Cell-free VZV virions are known to be produced in large quantities in skin lesions and are the type of particle that could be aerosolized and involved in viral transmission. In a study of leukemic recipients of live-attenuated varicella vaccine, only individuals with skin lesions as a side effect of varicella vaccination spread vaccine-type virus to varicella-susceptible close contacts. Similarly, the rare documented instances of transmission of vaccine virus from healthy vaccinees to other susceptible individuals (9 vaccinees who transmitted the virus to 11 contacts) occurred only when the vaccinee had a rash (including 4 cases from herpes zoster). Therefore the major source of infectious VZV seems to be the skin, although it is possible that transmission from the respiratory tract can also occur. Airborne spread of varicella has been documented, but indirect transfer by fomites has not. VZV DNA has been detected in air samples for many hours in hospitals, but the relationship to infectivity of the virus is unclear.

Incubation Period

The usual incubation period for varicella is 14 to 16 days. The range is 10 to 21 days, unless passive immunization has been given, in which case the incubation period may be prolonged to 1 week or more (usually up to 28 days).

Relationship Between Varicella and Zoster

It is documented that exposure of susceptible persons to zoster may result in varicella. Vesicular fluid from patients with zoster produced varicella when inoculated into susceptible children. Other studies have confirmed that a similar relationship exists under conditions of natural exposure. Claims to the contrary notwithstanding, it has not been documented that zoster is acquired from other patients with zoster or varicella. There are reports in the literature of occurrence of zoster in persons who happen to have been exposed to varicella or zoster, but these are chance events. The possibility that zoster is acquired from other patients with zoster of varicella was hypothesized in the past, but it is not supported by the current concepts of the pathogenesis of zoster, particularly the strict segmental distribution of lesions and the demonstrated presence of VZV DNA, RNA, and certain viral proteins in ganglia during latency. Studies have also determined that VZV DNA from zoster isolates is identical to that which caused the primary infection, proving that zoster is caused by reactivation of latent VZV.

Transplacental Transmission

In pregnancy, VZV may be transmitted across the placenta, resulting in congenital or neonatal varicella. The consequences of transplacental infection are discussed in a later section.

Epidemiology

Varicella is worldwide in distribution, and, in the absence of a vaccination program, it affects nearly every person by midadulthood. The epidemiology of varicella differs between temperate and tropical/subtropical regions. In temperate climates, infection occurs at younger ages (preschool aged children or children in early elementary school), with most children being infected by 15 years of age and less than 5% of adults remaining susceptible. There is a strong seasonal variation, with more cases and outbreaks occurring in winter and spring. In contrast, in tropical areas, children acquire varicella at older ages and a higher proportion of young adults remain susceptible, leading to a higher proportion of cases occurring among adults. The factors that determine the differences in epidemiology of varicella between temperate and tropical climates are not well understood but may relate to properties of VZV, which is thermolabile; climate influence other than temperature; population density; and the risk for exposure (e.g., differences in urban/rural residence and attendance in child care). Varicella is a more serious disease in young infants, adults, and immunocompromised persons, in whom there are higher rates of complications and deaths than in healthy children.

In the United States, before the introduction of live-attenuated varicella vaccine in 1995, varicella accounted for about 4 million cases, 11,000 to 15,000 hospitalizations, and 100 to 150 deaths every year. With the introduction of the routine one-dose regimen among children, significant changes in the epidemiology of varicella have occurred, with evidence of personal and herd immunity. There is less varicella disease occurring in all age groups, and the seasonality of the disease has disappeared. By 2005, varicella had declined approximately 90% compared with prevaccine years in active surveillance sites in which vaccination coverage had reached 90% among young children. The greatest decline (>90%) occurred in children aged 1 to 9 years, and an approximately 80% decline in incidence also occurred in infants not eligible for vaccination. Varicella-related hospitalizations declined greater than 75% to 88% from prevaccine years, and deaths decreased by 88% overall from 1990 to 1994 and from 2005 to 2007 ; in persons younger than 20 years, there was a 97% decline in deaths. Although the age-specific incidence has declined in all age groups, the median age at infection has increased, and cases occur predominantly in children in upper elementary school rather than in the preschool years. Seroprevalence studies in the early vaccine era (1999-2004) reflect the increased VZV seroprevalence among children (89% of children aged 6-11 years and 97% of those aged 12-19 years had VZV IgG antibodies) with maintenance of high seroprevalence among adults (98% of persons aged 20-49 years had VZV IgG). Cases continued to occur even in settings with high one-dose varicella vaccine coverage, and seropositivity did not always result after one dose. Therefore in 2006, a routine two-dose childhood varicella vaccination program with catch-up vaccination of all individuals without evidence of immunity was adopted in the United States. From 2006 to 2010, varicella incidence declined further (≈70%), and fewer outbreaks have been reported.

Varicella is more contagious than mumps but less so than measles. One study that compared all three diseases found that after exposure within households, 61% of susceptible persons of all age groups (without a history of previous disease) developed varicella, compared with 76% for measles and 31% for mumps. Another study found the attack rate for varicella at 87% among susceptible exposed household contacts. Compared with measles, varicella is about 80% as infectious in the household but only 35% to 65% as infectious in the community. The reason probably is that varicella requires closer contact for transmission, such as that occurring in the household, whereas in the community, there are more casual contacts. Measles may infect efficiently even through casual contacts.

Evidence of Immunity to Varicella

In the prevaccine era, a history of varicella was a valid measure of immunity. Because the rash is distinctive and subclinical cases occur rarely, most parents knew if their child had had varicella. Serologic testing has been used to assess the accuracy of reported histories of varicella. In adults, a positive history was highly predictive of serologic immunity; 97% to 99% of persons who reported a history were seropositive. However, the majority of adults who had negative or uncertain histories were also seropositive (71%-93%). This finding is supported by epidemiologic data from one study that found the attack rate after household exposure in parents who reported themselves as being susceptible was 5%.

In the early vaccine era, data indicated that the negative and positive predictive values of a history of varicella among adults were similar to those from the prevaccine era. In a study that included pregnant women attending prenatal care during 2001 to 2004 in Antelope Valley, California, and Philadelphia, who self-reported having had varicella, 98% had serologic evidence of immunity to varicella. Among those who reported a negative or uncertain history of varicella, 7% to 17% were seronegative. In another study, Perella and colleagues found similar predictive values among young adults (aged 15-29 years). However, this second study indicated that for cohorts born since 1994, the validity of reported varicella history is no longer highly predictive of seropositivity.

Considering the changes in the epidemiology of varicella in the United States, the Advisory Committee on Immunization Practices (ACIP) recommends as evidence of immunity a provider-verified history of disease (of varicella or zoster) rather than a self-reported history. Other criteria indicating evidence of immunity for adults include serologic evidence of immunity or laboratory confirmation of disease, documentation of receipt of two doses of varicella vaccine, and birth in the United States before 1980; however, birth in the United States before 1980 is not considered evidence of immunity for health care personnel, pregnant women, and immunocompromised persons. Table 23-1 presents evidence of immunity to varicella for pregnant women; the same criteria are used for varicella immunity for health care personnel and immunocompromised persons.

As more women of childbearing age have vaccine-induced immunity to varicella, the assessment of the immune status will shift toward documenting that the pregnant woman has received two doses of vaccine. For persons who have documentation of receipt of two doses, serologic testing to document presence of antibodies is not recommended because commercially available serologic assays for VZV IgG yield high rates of false-negative results.

Incidence of Varicella, Mumps, and Measles in Pregnancy

With varicella, mumps, and measles being highly infectious and primarily childhood diseases, a low proportion of adults, and inherently of women of childbearing age, are susceptible (<10%). Only a few studies have addressed the incidence of varicella, mumps, and measles during pregnancy; they were conducted before widespread use of vaccines and are not representative today. In a prospective study of clinically recognized infections that occurred during 30,059 pregnancies in 1958 to 1964, approximately 1600 women with presumed measles, varicella, and mumps were identified. This study’s findings suggested the minimum frequency per 10,000 pregnancies was 0.6 case for measles, 5 cases for varicella, and 10 cases for mumps. An estimate in 1992 projected an incidence of 7 cases of varicella per 10,000 pregnancies.

Today, measles and mumps are rare in the United States, and varicella is becoming less common. Given the high seroprevalence of antibodies either through vaccination or natural infection in the adult population and the decrease in circulation of measles, mumps, and VZV after implementation of the vaccination programs, varicella, mumps, and measles all are now unusual during pregnancy in the United States. It is likely that women immigrants from countries with tropical climates have a higher susceptibility to VZV and are at higher risk for contracting varicella during pregnancy than U.S.-born women if exposed; nonetheless, incidence is expected to be low because of low virus circulation.

Incidence and Distribution of Zoster

Zoster is primarily a disease of adults, especially older adults or immunosuppressed patients. Several studies in the United States found the incidence of zoster between 3.2 and 4.2 per 1000 population per year (age-adjusted to the 2000 population), translating into an estimated 1 million cases annually. This estimate was confirmed in the studies of Oxman on zoster vaccine. Hope-Simpson, describing patients of all ages in a general practice observed during a 16-year period in the United Kingdom, found an incidence of 3.4 cases per 1000 otherwise healthy people per year.

Adults and children older than 2 years who have zoster usually give a history of a previous episode of varicella; in younger infants with zoster, a history of intrauterine exposure to VZV can often be elicited. The latency period between primary infection and zoster is shorter if varicella occurs prenatally rather than later in childhood. Varicella in the first year of life increases the risk of childhood zoster, with a relative risk roughly between 3 and 21. Possibly, this phenomenon is caused by immaturity of the immune response to VZV in young infants, permitting early viral reactivation.

Age is the most important risk factor for the development of zoster. After infancy, the incidence of zoster increases with age, especially after 50 years of age (greater than two thirds of zoster cases occur in persons aged 50 years and older). The attack rate in octogenarians was 14 times that of children in the series by Hope-Simpson. Most studies have also shown that zoster incidence is higher among women than among men. Second attacks of zoster were historically considered unusual, although recent evidence suggests that the rates are comparable to those of the first episode in immunocompetent individuals. It is possible that some of the recurrent cases might be caused by reactivation of HSV; in one study, HSV was isolated from 13% of a series of 47 immunocompetent patients with clinically diagnosed zoster. Zoster in adults and children occurs with increased frequency in patients with altered cell-mediated immunity: malignant hematopoietic neoplasms (especially Hodgkin disease), patients after organ transplantation, and patients infected with human immunodeficiency virus (HIV). Spinal trauma, irradiation, and corticosteroid therapy may also be precipitating factors. The distribution of lesions in varicella, which primarily affects the trunk, head, and neck, is reflected in a proportionately greater representation of these regions in the segmental lesions of zoster.

Incidence of Zoster in Pregnancy

Recent studies on the incidence of zoster in pregnancy are lacking. In 1979, based on the number of live births, fertility rate, and incidence of zoster among persons aged 15 to 44 years, Brazin and colleagues projected approximately 6000 cases annually in pregnant women in the United States, which suggested that gestational zoster might be more common than gestational varicella. This is likely to change with the implementation of the varicella vaccination program. Assuming that there are 3.5 million pregnant women yearly in the United States, this calculates to a rate of 20 cases per 10,000 pregnant women per year. Nevertheless, zoster, similar to varicella, seems to be rare or uncommon in pregnancy. In contrast to varicella, little information exists regarding whether there is an increased severity of zoster in pregnant women compared with the general population. Therefore most experts infer that zoster in pregnancy is no more severe than it is in women who are not pregnant. Implications of gestational zoster for the fetus are discussed in a subsequent section.

Nosocomial Varicella in the Nursery

The precise risk of horizontal transmission in maternity wards or the newborn nursery after VZV has been introduced is unclear, but based on experience, it is very low. In reports in which the number of neonates exposed is stated, 249 exposures resulted in only 8 instances of transmission to infants; most of the exposed neonates had mothers with positive or uncertain histories of varicella. This low rate of transmission is likely due in part to the role that maternal immunity plays in protection of the infant and that, in general, most women of childbearing age in the United States are immune, with some higher susceptibility among women from tropical and subtropical areas outside of the United States. Additional factors supporting a low risk of transmission are relatively brief periods of exposure in health care settings compared with the household setting, where 80% to 90% of susceptible persons become clinically infected ; relative lack of intimacy of contact in the nursery, particularly for infants in isolettes; administration of postexposure prophylaxis; and high varicella immunity among health care workers because of prior disease or vaccination coupled with the requirement that health care workers have evidence of immunity to varicella.

Because IgG antibodies to VZV cross the placenta, full-term newborns of immune mothers are generally protected from varicella. Studies using sensitive assays to measure antibodies to VZV in the pre–varicella vaccine era demonstrated that high levels of passively acquired antibodies are present at birth and during the first 3 to 4 months of life; after 6 months of age, it was unusual for VZV antibodies to be detected. Antibodies to VZV have also been detected (>90%) in premature infants and infants with low birth weight born to mothers with a 25-week gestation or those weighing greater than 1000 g. Nevertheless, perinatal varicella has been reported rarely in infants born to women with positive histories of varicella or serology. In the study by Newman, varicella developed in a mother and her infant after exposure to a student midwife with varicella. The mother had experienced varicella as a child and had a few remaining skin scars; apparently, she had developed a second episode as an adult. Readett and McGibbon reported two cases of postpartum infection in neonates whose mothers had histories of varicella. After delivery at home, each of these infants was exposed within 24 hours of birth to a sibling with varicella and subsequently developed skin lesions when 12 and 14 days old. Their mothers did not develop varicella in the perinatal period and were found to have serum-neutralizing antibodies to VZV.

In the literature before 1975, VZV antibody titers were not often reported because sensitive tests for measuring these antibodies were not readily available. Since that time, infection of a few seropositive infants after postnatal exposure to VZV has been documented. These infants either had mothers with a history of varicella and VZV seropositivity or had VZV antibody that was transplacentally acquired. In one instance, mild varicella developed in a 2-week-old, 1040-g infant who was seropositive at exposure and was passively immunized with VZIG 72 hours after the exposure. In another study, five infants younger than 2 months, all of whom were seropositive at exposure, developed varicella in a children’s institution; all disease was mild (<20 lesions). Complete protection of every neonate against varicella is not guaranteed by immunity in the mother. However, when varicella develops in the presence of maternal antibodies, it seems to be modified.

Attenuated disease in young infants was also documented in a community-based study. In this study, clinical disease assessed by the proportion of infants with 50 or more lesions, fever, and varicella-related complications was milder among younger (0-5 months, n = 175) versus older (6-11 months, n = 344) infants, possibly because of the presence of maternal VZV antibodies. Eight infants were younger than 1 month at the time of varicella onset; all of them had fewer than 50 lesions, 1 infant was hospitalized, and 1 had a varicella-related complication (unspecified).

Reports largely confirm the low rate of transmissibility of varicella in neonates. Freud described an infant who had transplacentally acquired disease and developed lesions on the second day of life. None of the other neonates in the nursery became infected, but the index infant had been isolated immediately, so exposure had been very brief. When transferred to another ward, with older children, this same infant transmitted the disease to 2 older children, who were 4 and 7 years of age, respectively. Odessky and associates reported three instances of congenital varicella in which 2 infants were immediately isolated, but the third was not recognized as having varicella and exposed other neonates for 4 days. The number at risk is not stated, but no instances of transmission were observed. In a report by Harris, 35 infants were exposed to 2 infants with congenital varicella for periods of 18 and 10 hours, respectively, before isolation. None subsequently became infected, possibly because all the mothers had positive histories of varicella. In an additional case described by Matseoane and Abler, an infant developed transplacentally acquired varicella at 9 days of age and exposed 13 other neonates in the nursery for periods of 2 to 10 hours before isolation. Six mothers had a positive history of varicella, three did not, and four did not know. None of the exposed mothers or infants developed varicella. Friedman and colleagues report an outbreak in a neonatal intensive care unit (NICU) after a 25-week-gestation infant, whose mother had varicella 2 weeks previously, developed hemorrhagic varicella and exposed 14 infants over several days. None of the infants in isolettes became ill, but 4 who were in open warming units at exposure developed varicella 10 days later. All had received VZIG, and in each instance, the mother gave a history of varicella. The illnesses were mild, with only a few papular skin lesions, but 3 of the 4 infants were positive for VZV on immunofluorescence testing of skin scrapings. Each infant with varicella was treated with intravenous (IV) acyclovir. The incidence of disease was higher in infants of less than 29 weeks of gestation than in infants of longer gestation.

Lack of transmission despite hospital exposure to an adult with varicella in NICUs was also reported. In some reports, lack of transmission to the neonate is difficult to explain. In 1965, Newman reported two cases of varicella that occurred in mothers in the same prenatal ward 18 to 19 days after exposure to the index-infected infant and its mother. One mother developed varicella 7 days antepartum, and the other developed the disease 3 days postpartum. Each mother was immediately isolated from the ward but not from her own infant; neither of the infants developed varicella. In all, 139 mothers, excluding the index case, were exposed, and 8 developed infection. Three of 42 staff members also became infected. The index infant was the only neonate infected; all other infants, including those born to the 8 infected mothers, remained free of disease. ( Experiences with nosocomial varicella in the newborn nursery with additional details were presented in the previous editions of this textbook [see Table 22-1 in the 2010, seventh edition] ).

Since introduction of the varicella vaccine in the United States, one experience describing exposure in a neonatal unit has been reported. A neonatal nurse developed varicella after being exposed to her son, who had breakthrough varicella despite a history of two doses of varicella vaccine. The nurse exposed 34 neonates in the NICU. All exposed neonates were given IV immunoglobulin, and those born at less than 28 weeks of gestation (9) also received acyclovir; none developed varicella.

The varicella vaccination program implemented in the United States in 1995 greatly impacted both VZV circulation in the community and the profile of population immunity, which influenced nosocomial varicella, including the nursery. A community-based study found that the incidence of varicella in infants in the United States declined 90% from 1995 to 2008. A similar experience was described from Australia, where there has been a reduction in the frequency of congenital varicella syndrome (100%) and neonatal varicella (85%) in the first years after the introduction of the universal varicella vaccination program. With high population immunity, the risk of exposure will be lowered further. In terms of population immunity, more vaccinated women will enter childbearing age. Similarly to measles, the amount and persistence of VZV maternal antibodies transferred to infants of vaccinated women are likely to be lower than those transferred from naturally infected women. Therefore infants born to vaccinated mothers are likely to have lower levels or more rapid declines of maternal antibodies and may be at an increased risk of varicella and related complications if exposed and infected in the early months of life. Even if low, the risk for nosocomial transmission of varicella in nurseries remains, and measures to prevent transmission should be immediately implemented, considering the potential for severe disease in this population. Maintaining high vaccination coverage among health care personnel caring for these infants and among their household contacts will increase in importance.

Cutaneous Lesions

Histologic changes in the skin leading to the formation of vesicles are essentially identical for varicella, zoster, and HSV infection. The hallmark of each is the presence of multinucleated giant cells and intranuclear inclusions, changes that are not found in the vesicular lesions caused by vaccinia virus and coxsackieviruses. The lesion is primarily localized in the epidermis, where ballooning degeneration of cells in the deeper layers is accompanied by intercellular edema. As edema progresses, the cornified layers are separated from the more basal layers to form a delicate vesicle with a thin roof. An exudate consisting primarily of mononuclear cells is seen in the dermis, but the characteristic nuclear changes of epithelial cells are absent in this region.

The predominant cells in vesicular lesions are polymorphonuclear leukocytes. These cells may play a role in generating interferon in vesicular lesions, which may be important in recovery from the disease. In vitro data also suggest that the polymorphonuclear leukocyte plays a role in host defense against VZV, possibly by mediating antibody-dependent cell-mediated cytotoxicity. Cytotoxic T cells play a role in recovery from VZV infections, and recent immunologic information on children with severe varicella or zoster has also implicated natural killer (NK) cells in host defense against the virus.

Visceral Lesions in the Fetus and Placenta

Few reports describe the appearance of the placenta in cases of congenital varicella with or without survival. Garcia observed grossly visible necrotic lesions of the placenta in a case of varicella occurring in the fourth month of pregnancy that resulted in spontaneous abortion. Microscopically, central areas of necrosis were surrounded by epithelioid cells and rare giant cells of the foreign body type, giving a granulomatous appearance. Some decidual cells had typical intranuclear inclusions.

Descriptions of the pathology of visceral lesions in fetal or neonatal varicella are restricted to autopsies in fatal cases. Grossly, the lesions are small, punctate, and white or yellow and resemble miliary tuberculosis. Microscopically, their appearance resembles the lesions of the placenta: central necrotic areas, often resembling fibrinoid necrosis, surrounded by a few epithelioid cells and a scant infiltrate of mononuclear cells. Intranuclear inclusions are present. The skin, lungs, and liver are uniformly involved ( Table 23-2 ). In the case described by Garcia, the cortical, subependymal, and basilar structures of the cerebrum were totally destroyed and accompanied by extensive calcification. Although a search for Toxoplasma was negative, serologic data to rule out dual infection are lacking in the report. The gross and microscopic lesions of fatal perinatal varicella resemble lesions of disseminated HSV infection, including a preference for the liver and adrenal gland, but the provided data suggest that involvement of the brain is more common in neonatal HSV infection than it is in fatal neonatal varicella. A neonate with fatal hemorrhagic varicella with pneumonia and hepatitis is shown in Figure 23-1 .

Visceral Lesions in the Mother

In fatal cases of varicella in pregnant women, maternal death is usually caused by pulmonary involvement. The pathologic course of varicella pneumonia in pregnant women is identical to the course in nonpregnant women and in children. Interstitial pneumonitis may follow a peribronchiolar distribution of disease. Intranuclear inclusions may be found in alveolar lining cells, macrophages, capillary endothelium, and tracheobronchial mucosa. Necrotic foci may be accompanied by hemorrhage, and hyaline membranes lining the alveoli are often prominent.

Zoster

The pathologic picture of cutaneous lesions in zoster is indistinguishable from varicella lesions. The dorsal root ganglion of the affected dermatome exhibits a mononuclear inflammatory infiltrate. There may also be necrosis of ganglion cells and demyelination of the corresponding axon. There are no descriptions of these lesions in pregnant women or in neonates specifically.

Varicella Rash

After an incubation period of usually 13 to 17 days, varicella is heralded by the approximately simultaneous occurrence of fever and rash. In adults, the exanthem is often preceded by a prodromal fever and constitutional symptoms lasting 2 or 3 days. On occasion, one or more isolated vesicles may precede a generalized exanthem by 1 or 2 days. The rash is characteristically centripetal, beginning on the face or scalp and spreading rapidly to the trunk but with relative sparing of the extremities. The lesions begin as red macules but progress quickly to vesicles and crusts. Itching is the rule. There is a tendency for new lesions to occur in crops. In contrast to smallpox, all stages of lesions—vesicles, pustules, and scabs—may occur simultaneously in the same anatomic region. New crops often continue to appear over a 2- to 5-day period. Lesions may be more numerous in skin folds or in the diaper area. The total number of vesicles varies from only two or three in very mild cases, especially in infants, to thousands of lesions that border on confluence, especially in adults. In many cases, one or two mucosal lesions may occur in the mouth or, less commonly, on the vulva. On occasion, the lesions may be bullous or hemorrhagic. Residual scarring is exceptional. Constitutional symptoms tend to be mild, even in the presence of an extensive exanthem.

Complications of Varicella

The most common complication is secondary bacterial infection, usually caused by group A β-hemolytic streptococci or staphylococci. Skin infections may lead to severe sequelae, such as toxic shock syndrome and necrotizing fasciitis. Central nervous system (CNS) complications, which are uncommon, include encephalitis, cerebellar ataxia, aseptic meningitis, stroke, and vasculopathy. Glomerulonephritis, myocarditis, and arthritis have also been reported.

Varicella in Immunocompromised Children

It is widely appreciated that varicella may be severe and even fatal in children with an underlying malignancy, children with congenital deficits in cellular immunity, children receiving high doses of corticosteroids for any reason, and children with underlying infection with HIV and acquired immunodeficiency syndrome (AIDS). Historically, children with leukemia had a mortality rate approaching 10% if untreated and sometimes developed what has been called progressive varicella. Instead of developing new vesicular lesions for several days, they continued to have fever and new lesions for 2 weeks after the onset of illness. Frequently, their skin lesions become hemorrhagic, large, and umbilicated. Varicella pneumonia often ensued and was a major factor contributing to the death of a child. It is believed that this abnormal response to VZV represents a failure of the normal cell-mediated immunity response to eliminate the virus. The cell-mediated immunity response to VZV includes antibody-dependent cell-mediated cytotoxicity; NK cells; cytotoxic T cells, including CD4 and CD8 cells; and NK cells. Today, in the vaccine era, deaths from varicella have become rare.

Varicella Pneumonia

Primary varicella pneumonia is a dreaded complication of varicella and is responsible for most fatalities. It is most common in immunocompromised patients, in adults, and in most cases of fatal neonatal varicella, but it is rarely seen in otherwise healthy children. It has been suggested that the incidence is about 15% in adults and that 90% of cases have occurred in persons older than 19 years. The true incidence is difficult to determine because chest radiographs are not performed in most cases of varicella, and extensive radiographic evidence of disease may be present when pulmonary symptoms are only minimal. In male military recruits with varicella, virtually all of whom had been hospitalized and had chest radiographs, radiographic evidence of pneumonia was found in 16.3% of 110 cases.

Two reviews of varicella pneumonia in adults outline the major features. The onset of pneumonia usually occurs in 2 to 4 days but sometimes occurs 10 days after the appearance of the exanthem. Fever and cough are present in 87% to 100% of cases, and dyspnea occurs in 70% to 80%. Other symptoms and signs include cyanosis (42%-55%), rales (55%), hemoptysis (35%-38%), and chest pain (21%). Radiographic changes seem to correlate best with the severity of the rash rather than with the physical examination of the lungs. The radiograph typically reveals a diffuse nodular or miliary pattern, most pronounced in the perihilar regions. The radiographic appearance changes rapidly. The white blood cell count ranges from 5000 to 20,000 cells/mm ³ and is of little help in differentiating viral from secondary bacterial pneumonia. Pneumonia is usually self-limiting, and recovery is temporally correlated with clearing of skin lesions. The fatality rate has been variously estimated at 10% to 30%, but it probably approximates the lower of these values if immunocompromised hosts are excluded. Blood gas analyses and pulmonary function tests indicate a significant diffusion defect that may persist in some cases for months after clinical recovery. The introduction of antiviral chemotherapy has greatly improved the outcome in this disease.

Maternal Effects of Varicella

Reports from the mid-20th century suggested that when varicella occurred during pregnancy, it was a highly lethal disease. Deaths usually resulted from varicella pneumonia, in some cases accompanied by glomerulitis and renal failure or myocarditis, occurring after the fourth month of gestation. Harris and Rhoades reviewed the literature to 1963 and found a reported mortality of 41% for 17 pregnant women with varicella pneumonia compared with 11% for 236 nonpregnant adults with varicella pneumonia. Other reports question, however, whether varicella, especially in the absence of pneumonia, is more serious in pregnant women than in the adult population at large. Because most cases of gestational varicella with an uncomplicated course are undoubtedly not reported, the denominator of the case-fatality ratio (CFR) is unknown. In a prospective study of 150 cases of varicella in pregnancy in 1966, only one maternal death related to varicella pneumonia was recorded.

In a very large, collaborative, prospective study published in 2002, there were no fatalities in 347 consecutive pregnant women with varicella, although 18 (5.2%) had radiologic evidence of pneumonia. Although the data did not reach statistical significance in this study, it seems striking that 16 (89%) of 18 reported cases of pneumonia occurred in women who developed varicella after the 16th week of pregnancy.

Based on case reports and reviews of gestational varicella (with and without pneumonia), 542 cases of varicella in pregnant women have been reported since 1963. There were 16 deaths (3%). All of the deaths occurred among the 75 (14%) women who had varicella pneumonia (21% fatality rate for pneumonia). Deaths occurred in 1 (<1%) of 166 women whose disease occurred during the first trimester, 4 (2%) of 168 women whose disease occurred during the second trimester, and 11 (5%) of 208 women whose disease occurred during the third trimester. Antiviral therapy was used only after 1985. No deaths occurred among an additional 8 women who were exposed to varicella in late pregnancy but did not develop an exanthem until the first few days postpartum. ^∗

∗ References .

It remains uncertain whether varicella pneumonia has a graver prognosis when it occurs during pregnancy. There is no definitive evidence that varicella in the absence of pneumonia is a more serious illness in pregnant women than in other adults; however, the risk of developing pneumonia may be increased after the 16th week of pregnancy. It seems likely that older mortality information on varicella in pregnancy reflected the pre–antiviral therapy era and was biased by selective reporting of fatal cases.

Some patients with varicella during pregnancy who were treated with acyclovir have been reported. These reports suggest that acyclovir has improved the outcome of this complication of varicella, although controlled studies have not been performed. Although various dosages have been used, the standard dosage of 30 mg/kg/day given IV would seem appropriate for treatment of pregnant women with varicella pneumonia. Congenital abnormalities from administration of acyclovir to women during pregnancy have not been observed.

Controlled studies of the value of corticosteroids in pregnant women with varicella pneumonia have not been performed. Several reports indicate that 2 of 6 pregnant women treated with corticosteroids died, whereas 8 of 17 pregnant women given supportive therapy without corticosteroids died. It seems that administration of an antiviral drug is of greater importance than administration of corticosteroids. Passive immunization may be administered to seronegative women after close exposure to VZV to attempt to modify the infection; although uncertain, this approach may prevent fetal infection. In a study from 1994, among 97 women who developed varicella after passive immunization with VZIG, there were no observed cases of congenital varicella syndrome. About two abnormal infants could be expected in a series of this magnitude, but the number of women followed is too low to achieve statistical significance.

Effects of Gestational Varicella on the Fetus

Congenital Malformations

For many years, there was uncertainty about whether gestational varicella led to a symptomatic congenital infection. Intensive investigation from the mid-1970s until the end of the 20th century led to the recognition that VZV can cause fetal malformations. Two types of investigations were done to determine whether varicella during pregnancy leads to a congenital syndrome. The first investigations were retrospective analyses or case reports describing specific anomalies that occurred in the offspring of mothers who had gestational varicella. These reports were necessarily highly selective and did not define the incidence of such anomalies. They consistently described a syndrome of skin scarring, eye and brain damage, and limb hypoplasia, however, that might follow intrauterine varicella.

The second type of analysis consisted of prospective studies of pregnant women followed throughout pregnancy and afterward. The problem was to delineate the coincidence of two events, each of which is itself uncommon—gestational varicella and congenital malformations—to determine the magnitude of risk to the fetus. Siegel, despite an 8-year observation period encompassing approximately 190,000 pregnancies annually in New York City, was able to identify only four malformations among infants born to 135 mothers who had varicella during pregnancy, compared with five malformations among 146 matched controls. The follow-up period was 5 years and included psychomotor and audiometric tests. Varicella occurred during the first trimester in only 27 of the pregnancies complicated by varicella, and of these, 2 (7.4%) were associated with congenital anomalies compared with anomalies in 3 (3.4%) of 87 pregnancies in the control population.

The largest single prospective series is that of Enders and associates. In a joint prospective study in Germany and the United Kingdom from 1980 to 1993, Enders and associates followed 1373 women with varicella and 366 with zoster during pregnancy. Of the women with varicella, 1285 continued to term, and 9 infants had defects attributed to congenital varicella syndrome. The incidence was 2 (0.4%) of 472 for infections between 0 and 12 weeks and 7 (2%) of 351 for infections between 13 and 20 weeks. In a collaborative prospective study in the United States, 347 women with gestational varicella were reported, and adequate follow-up of their infants was available in 231. In this cohort, there was one case (0.4%) of the congenital syndrome and two cases of fetal death, including one case of hydrops. If these cases are included, the rate of congenital varicella was 1.3%. The mother of the one child with the syndrome had varicella at 24 weeks; the child had skin, eye, and CNS involvement.

That congenital varicella syndrome is a reality is now widely appreciated. It has become possible to make a tissue diagnosis of congenital varicella syndrome only more recently because affected infants do not chronically shed virus as is seen in congenital infections with rubella virus and CMV. Congenital varicella syndrome may be prevented in the future by widespread use of varicella vaccine, analogous to the situation for congenital rubella.

The constellation of developmental abnormalities described in individual case reports of infants born to mothers who had varicella in early pregnancy and in prospective series is sufficiently distinctive to indicate that VZV is a teratogen. In 1947, LaForet and Lynch described an infant with multiple congenital anomalies after maternal varicella in early pregnancy. The infant had hypoplasia of the entire right lower extremity, talipes equinovarus, and absent deep tendon reflexes on the right. Cerebral cortical atrophy, cerebellar aplasia, chorioretinitis, right torticollis, insufficiency of the anal and vesical sphincters, and cicatricial cutaneous lesions of the left lower extremity were present. The syndrome then seemed to be all but forgotten until 1974, when Srabstein and coworkers rekindled interest in the subject by reporting another case and reviewing the literature, concluding that although the virus could not be isolated from the infants, congenital varicella syndrome typically consisted of some combination of cicatricial skin lesions, ocular abnormalities, limb deformities, mental retardation, and early death after maternal varicella in early pregnancy ( Figs. 23-2 to 23-4 ). Numerous additional reports in the literature of the syndrome, encompassing more than 100 cases, indicate there is a wide spectrum of manifestations ( Table 23-5 ). ^‡

‡ References .

Table 23-5

Reported Symptoms and Signs in Infants With Congenital Varicella Syndrome, 1947-2002

See references

Symptom	Estimated Incidence (%)
Skin lesions (cicatricial scars, skin loss)	60-70
Ocular abnormalities (chorioretinitis, Horner syndrome, anisocoria, microphthalmia, cataract, nystagmus)	60
Neurologic abnormalities (cortical atrophy, mental retardation, microcephaly, seizures, dysphagia, limb paresis)	60
Abnormal limbs (hypoplasia, equinovarus, abnormal or absent digits)	50
Prematurity, low birth weight	35
Death in early infancy	25
Abnormalities of gastrointestinal tract	10
Urinary tract abnormalities	10
Zoster in infancy	20

Although at one time it was thought that congenital varicella syndrome occurred after maternal VZV infection in the first trimester of pregnancy, current evaluation of the data indicates that cases also occur in the second trimester. Of 82 cases for which data are available, 32 (39%) occurred after maternal varicella that developed before week 13, 47 (59%) occurred after maternal varicella that developed between weeks 13 and 26, and 1 (1%) occurred after maternal varicella that developed during week 28. The average gestation when maternal varicella occurred was 15 weeks. Only 6 cases occurring after maternal zoster have been reported ; not all of these are well documented virologically. Four occurred after maternal zoster in the first trimester, one followed zoster in the second trimester, and one followed zoster in the third trimester. Of 109 reported affected infants, 103 (95%) cases followed maternal varicella, and 6 (5%) followed maternal zoster (disseminated in one instance).

Scars of the skin, usually cicatricial lesions, are the most prominent stigmata, although a few patients have had no rash at all. Eye abnormalities (i.e., chorioretinitis, microphthalmia, Horner syndrome, cataract, and nystagmus) and neurologic damage are almost as common; other features include a hypoplastic limb, prematurity, and early death. The features of the syndrome are summarized in Table 23-5 .

Cutaneous scars were usually observed overlying a hypoplastic limb but also have been seen in the contralateral limb. Characteristically, the skin scars are cicatricial, depressed, and pigmented and often have a zigzag configuration. Such scars are thought to be the result of zoster that occurred before birth. In some patients, large areas of scarred skin have required skin grafting. In other patients, the rash was bullous or consisted of multiple, scattered, depressed, white scars. In one infant, healing zoster was present at the T11 dermatome at birth; there was also spinal cord atrophy at the same level and aganglionosis of the intestine.

Ocular abnormalities include chorioretinitis, Horner syndrome or anisocoria, microphthalmia, cataract, and nystagmus. ^§

§ References .

Rarely, major abnormalities were confined to the eye. There was no apparent effect of timing of maternal varicella during gestation; the times of infection varied from 9 to 23 weeks in these infants. Figure 23-2 is a photograph showing retinal involvement in one of these patients.

Neurologic involvement is about as common as skin and eye abnormalities in infants with congenital varicella syndrome. Patients with cerebral cortical atrophy, diffuse brain involvement, or mental retardation (frequently accompanied by abnormal electroencephalograms and seizures or myoclonic jerks) have been described. In a few patients, cerebrospinal fluid (CSF) findings were normal ; in others, there were increased numbers of leukocytes or protein levels. Bulbar palsy is suspected to result in dysphagia and bouts of aspiration pneumonia in some of these children. ^‖

‖ References .

Deep tendon reflexes were reported as normal in one infant and diminished to absent in six, ^¶

¶ References .

and they were in some cases accompanied by sensory deficits. Electromyography in some patients revealed a denervation pattern with loss of motor units. A biopsy specimen in one instance showed replacement of muscle bundles by fat. At least five children with vocal cord paralysis have been reported.

Abnormalities of the limbs can be extremely dramatic in presentation and are seen in about half of affected infants. The most common limb abnormality, which first called attention to this congenital syndrome, is hypoplasia of a limb, most commonly unilateral involvement of a leg or arm (see Table 23-5 ). Hypoplasia or absence of digits has also been observed. Talipes equinovarus or a calcaneovalgus deformity has also occurred. ^#

# References .

This complex of abnormalities in the limbs, including the bony abnormalities, is probably attributable to a neuropathy caused by direct viral invasion of the ganglia and spinal cord.

About one fourth of these infants died within the first 14 months of life. One infant with the obvious syndrome was stillborn. In one infant who died at 6 months, autopsy revealed a necrotizing encephalitis with various degrees of gliosis and inflammatory infiltrates. Focal calcification was observed in white and gray matter of the cerebrum, brainstem, and cerebellum. Atrophy of the anterior columns of the spinal cord and scarring in the ganglion corresponding to the distribution of the skin lesions and an atrophic limb were also present. No inclusion bodies were identified. Among infants with a hypoplastic limb, 40% had evidence of mental retardation or died early. The presence of a hypoplastic limb on an ultrasound examination suggests a poor outcome.

About one third of affected infants were premature or had low birth weight for their gestational ages, and about 10% had various abnormalities of the gastrointestinal tract, including reflux, duodenal stenosis, jejunal dilation, microcolon, atresia of the sigmoid colon, and sphincter malfunction. ^∗a

∗a References .

A similar percentage had abnormalities of the urinary tract, often caused by poor or absent bladder sphincter function. ^†a

†a References .

Involvement of the cervical or lumbar spinal cord and the autonomic nervous system is thought to account for the observed hypoplasia or aplasia of limbs and digits, motor and sensory defects, decrease or absence of deep tendon reflexes, Horner syndrome, and gastrointestinal and urinary tract abnormalities.

Figures 23-3 and 23-4 depict two children with stigmata of congenital varicella syndrome. One has severe and one has relatively mild involvement.

zoster after congenital varicella syndrome

Of children with congenital varicella syndrome, 15% develop clinical zoster in infancy or early childhood, almost all in the first year of life. ^‡a

‡a References .

This finding is of particular interest because cell-mediated immunity to VZV in 2 of 10 of children with the syndrome has been reported to be absent as determined by lymphocyte transformation. In the series by Enders and associates of 1291 live births (without congenital varicella syndrome), of whom conservatively perhaps 25% were infected with VZV (the attack rate could be as high as 50%), the rate of zoster in childhood was 3%. Zoster seems to be more common in children with congenital varicella syndrome than in infants who were infected with VZV in utero but were asymptomatic at birth.

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here